The most critical challenge facing humanity today is the radical improvement of energy efficiency in neural networks. Current approaches — where increasing computational power requires building ever larger power plants to sustain artificial intelligence systems — represent a dead-end path.
Our task is to find a fundamentally different trajectory, based on new materials, architectures, algorithms, and physical principles.
The Center for Neurophysics and Neuromorphic Technologies conducts research into the fundamental foundations of such solutions. In particular, we study “smart” materials that can serve as the building blocks for energy-efficient computing systems, and we investigate the architecture of complex networks — from brain structures to social interactions — in order to uncover universal principles underlying efficient systems, among other directions.
Nature already provides a compelling example: the human brain, which combines immense computational power with minimal energy consumption. This demonstrates that the problem is, in principle, solvable — and our goal is to translate these principles into the technologies of the future.
